Explain why each group uses a dedicated node
Pair: @kjnilsson.
This commit is contained in:
parent
d1e8ac6c8e
commit
ca3612ba62
|
|
@ -174,6 +174,12 @@ validation_success_for_AMQP_client(Config) ->
|
|||
?MODULE, validation_success_for_AMQP_client1, [Config]).
|
||||
|
||||
validation_success_for_AMQP_client1(Config) ->
|
||||
%% This test intentionally doesn't whitelist any certificates.
|
||||
%% Both the client and the server use certificate/key pairs signed by
|
||||
%% the same root CA. This exercises a verify_fun clause that no ther tests hit.
|
||||
%% Note that when this test is executed together with the HTTP provider group
|
||||
%% it runs into unexpected interference and fails, even if TLS app PEM cache is force
|
||||
%% cleared. That's why originally each group was made to use a separate node.
|
||||
AuthorityInfo = {Root, _AuthorityKey} = erl_make_certs:make_cert([{key, dsa}]),
|
||||
{Certificate, Key} = chain(AuthorityInfo),
|
||||
{Certificate2, Key2} = chain(AuthorityInfo),
|
||||
|
|
|
|||
Loading…
Reference in New Issue